The technology in healthcare is evolving at a rapid pace, and the healthcare industry is felicitous for significant changes. The sudden technological advancements have uprooted the traditional working setting of the medical community, showing frustration and dissatisfaction in the short term. Physician burnout crisis has become a prominent challenge in the healthcare industry. Healthcare providers are experiencing overwhelming frustration and a loss of professional fulfilment. [1]

Also, we are still battling with pathogens which have been with us since the advent of human history such as tuberculosis and novel pathogens from non-human hosts which are posing a major threat to the public health. The emergence of new viral infections within the last ten years that has potentially spread across international borders; the latest of this being is novel coronavirus (COVID-19) has to wreak global havoc. [2,3]

 

There is an urgent need to improve the staggering levels of physician burnout around the globe and develop responsible, data-driven solutions to help us through the current worsening humanitarian crisis in the long run. Endless opportunities are available to introduce technology in healthcare for developing more precise, practical and impactful interventions inpatient care, from chronic disease and cancer to COVID-19. The development of the recent machine learning-based tools and techniques has provided novel ways to combat such chronic disease and global pandemics. [1-3]

What is Artificial intelligence (AI)?

 

Artificial intelligence is a computer system able to perform tasks that ordinarily require human intelligence. AI has been found to be a potential engine to drive improvements across patient care. It can be defined as a conglomeration of technologies and concepts that resemble processes correlated with human intelligence such as learning, adaptation, sensory understanding and interaction. Development of data analytics and increased healthcare data availability has lead to the recent successful applications of AI in healthcare and medicine.

AI techniques guided by relevant clinical questions can help to collect clinically relevant information from a higher amount of data, which will further assist in decision making or even replace human judgement in specific functional areas of healthcare. It can be used to improve physicians workflow while providing burnout relief and reducing burden. [4]

EFFECTS OF AI IN HEALTHCARE AND MEDICINE

1) Reduce the burden of Electronic health records (EHR):

Electronic health records (EHR) plays an essential role in the healthcare industry towards, but it has been associated with other problems such as user burnout and endless documentation.  

AI can be used to automate some of the routine processes that consume greater user time. It can also be used to process routine requests from the inbox and scheduling tasks that immediately requires the clinician's attention. Voice recognition and dictation can help to improve the clinical documentation process.

2) More precise analytics for pathology and diagnosis:

Pathologists provide significant diagnostic data for providers. Most of the decisions in healthcare are based on the pathology; therefore, an increase in accuracy will increase the efficiency of diagnosis.  

 

Digital pathology and AI analytics can help to improve pixel level on extremely large digital images allowing pathologists to identify nuances which may escape the human eye. It can also improve productivity by identifying features of interest in slides, thereby increasing the efficiency and reducing the time needed for diagnosis of each patient. 

3) Transform healthcare data into actionable insights:

 

The data entered into electronic health records (EHR) needs to be processed in any discernible way. Computers and AI helps to identify data patterns and determines which data is impactful for outcomes. It allows health care organizations to generate patient summaries in a concise and meaningful way and generates new information in real-time.  

AI technology helps to decipher this information and can evaluate its importance in decision making. It can also provide new insights derived from the conversation and identify potential gaps in care and differential diagnosis.

4) Brings intelligence to medical devices:

Smart devices are critical for monitoring the patients in the ICU. Assessment of complications and sepsis risk can significantly improve outcomes and reduces healthcare cost.  

 

Artificial intelligence can integrate disparate data from across healthcare system and enhance the ability to identify complications- something that a human can't do well. Inserting intelligent algorithms into these devices also reduce the cognitive burdens of physicians with patient assurance in less time.

5) Makes cell phones a powerful diagnostic tool:

Experts believe that imaging using smartphones will be an essential supplement to clinical quality imaging. The increased quality of cell phone images is viable for analysis by artificial algorithms.

 

Images of eyes, infections, wounds, medications or other subjects collected from smartphones might help to cope with unreserved areas with a shortage of specialists in a short time. The areas of dermatology and ophthalmology are already getting the benefit of this trend.

6) Improves drug development, making it fast:

Developing a drug is a complicated, time taking and expensive procedure. The drug development processes can be improved using machine-based learning algorithms and AI tools. It can help in analysing available data to identify potential target molecules. Millions of the potential target molecules identified can be filtered down to select the most appropriate and suitable molecule depending on the structural fingerprints and adverse outcomes.

It can also speed up the process of drug development and clinical trials, by choosing relevant therapeutic molecules from the libraries with the help of automatic machine learning algorithms. This can serve as a warning sign for the clinical trials that are not producing conclusive results, therefore allowing researchers to intervene earlier, saving both time and money in drug developments. It can help to identify potential biomarkers which help in more accurate diagnosis of disease compared to complicated procedures such as whole-genome sequencing.

7) Identifies triage tools for emergency:

It is integral to identify the severity of a patient medical condition for early identification of vulnerable and high-risk populations, especially in emergency care units.

 

Online triage tools developed and validated using A.I. based algorithm and deep learning can help in accurate prediction of critical care in patients compared to conventional triage tools. AI-based technology can help to facilitate triaging before the patient reach point of care.

8) Upgrades gene editing:

Clustered Regularly Interspaced Short Palindromic Repeats (CRISPR), especially the CRISPR-Cas9 system is the widely used gene-editing system, providing a leap forward in DNA editing with both cost-effective and high precision. This gene-editing approach is based on targeting and editing a gene locus using short guide RNAs (sgRNA). The sgRNA can fit in multiple locations and can cause unintentional adverse effects.

 

Careful selection of sgRNA with least adverse events is of utmost importance in the application of CRISPR-CAS technology. Machine learning algorithms and deep learning can help to provide the best outcomes in predicting both guide-target interactions and off-target effects for a specific sgRNA. This can speed up the process of sgRNA development for every region of human DNA segments. [5]

BENEFITS OF AI AMIDST THE COVID 19 PANDEMIC

1) Helps in outbreak detection

• Machine learning, analytics, and natural language processing (NLP) are used in biosurveillance. Scanning of social media, news reports, and other online data can be useful in localizing disease outbreaks before they reach the level of the pandemic.

• Satellite imaging and big data analysis has been used successfully in the past to detect outbreaks.

• Sentiment analysis can help understand the public's reactions or overreactions towards disease outbreaks, and therefore can provide valuable insights to the government in directing efforts towards public education.

For instance: The Canadian company Blue Dot had successfully used machine learning algorithms to detect early outbreaks of COVID-19 in Wuhan, China by the end of December 2019.

2) Predicts the spread of infection

• Dynamic, statistical and mathematical predictive models have been utilised for predicting the spread and extent of infectious diseases.

• Traditionally used predictive models provides trend-based recalibration, adaptive learning, scope and flexibility to improve prediction.

• These models provide a new understanding of the disease process and help in estimating the impact of the interventions, such as social distancing, in preventing the spread of infection.

For instance: The Susceptible-Exposed-Infectious-Recovered (SEIR) modelling method is now used for predicting the areas and extent of COVID-19 spread. By evaluating parameters of the epidemic, such as under-reporting of cases, the effectiveness of interventions, and the accuracy of testing methods

3) Formulate preventive strategies and helps in vaccine development

• Artificial Neural Networks (ANN) had been used for predicting antigenic regions with a high density of binders in the viral membrane protein of Severe Acute Respiratory Syndrome Coronavirus (SARS–CoV).

• Machine learning tools helps in rapid scanning of the entire viral proteome, allowing faster and cheaper vaccine development.

• Machine learning tools can be used to predict the hosts of newly discovered viruses based on nucleoprotein gene sequences and spike gene sequence analysis

• It is useful for tracing back viral origins, especially when the data set is large and comparative analysis is difficult or time-consuming

For instance: Prediction of the future expansion of small subtrees of hemagglutinins (HA) part of the viral antigenic set was possible from training H3N2 and testing on H1N1, using reconstructed timed phylogenetic tree

4) Useful in early case detection and tracking

• Machine learning tools can be used to collect, analyze large amounts of data stratify patients as per risk and propose treatment solutions.

• Quarantining, early case identification and prevention of exposure to the communities are the important pillars for management of an epidemic such as COVID-19.

• Mobile phone-based surveys are useful in early identification of infected cases in quarantined populations.

• Digital phenotyping is the novel method for collecting smartphone-based active and passive data to produce an individual phenotype.

For instance: Government of India has recently launched a mobile application "Aarogya Setu" which tracks user exposure to potentially COVID-19 infected patients, using the Bluetooth functionality to scan the surrounding area for other smartphone users.

COVID-Net, a deep CNN design can be used for detection of COVID-19 cases from CT images and X rays.

5) Prognosis prediction

• Machine learning algorithms are widely used for prognosis of patients being infected with MERS Co-V infection.

• The major predictors in the recovery of patients are age, disease severity and the presence of pre-existing co-morbidities.

• The machine learning-based CT radiomics models showed feasibility and accuracy for predicting hospital stay in COVID-19 patients

For instance: Supervised XGBoost classifier provides a simple and intuitive clinical test to precisely and quickly quantify the risk of death.

6) Development of treatment approaches

• Machine learning tools have been utilised in drug development and enable to interpret large gene expression profile data to suggest novel uses of currently available medications.

• Deep generation models, also called as AI imagination, can help in the formulation of novel therapeutic agents with desirable outcomes. It will reduce the time and cost of drug development while improving diagnosis and discovering novel therapeutic agents.

For instance: Adversarial autoencoders is used to disentangle the style and content of images, unsupervised clustering, dimensionality reduction and data visualization.

 

Below is an example of machine learning model developed for COVID-19, which helps in identifying outbreak, prediction models for spread, prevention and vaccine development, early case detection and tracking, prognosis prediction of affected patients, and drug development. (Fig.2) [6,7]

BENEFITS OF AI IN HEALTHCARE

The major benefits that AI provides to deliver better outcomes in health care are as follow:

 

1. Enhances the quality of healthcare delivery with better responses compared to human intelligence

2. Increases clinical productivity by handling large amounts of data with better efficiency to predict patient outcomes

3. Improves decision making by analyzing patient record, symptoms and data structure rather than medical care based on intuitions and predictions

4. Provides higher levels of accuracy with minimized risk without any burnout

5. Delivers quality health assistance and precision medications by improving in-person or online consultations, diagnostics, and treatment plans

6. Expands access to care in under-reserved and developing regions by mitigating the shortage of qualified clinical staff and trained healthcare providers

7. Reduces mortality by early detection of high-risk diseases and efficient medical management

8. Improves health care cost and outcomes by integrating information, creating time-saving administrative duties and reducing hospital visits

 

 

PITFALLS OF USING AI 

MAJOR CHALLENGES OR BARRIERS TO UTILISE AI

 

Various instances of AI do exist in health care in the present scenario, but there are some major challenges faced by AI developers and medical opportunity in making AI a mainstream in medicine. These challenges are as follows:

Positive ongoing efforts are needed for solving the above challenges so that AI algorithms can become a common part of medical practices in future. [8]

 

REMARKABLE AND REVOLUTIONARY AI INNOVATIONS IN HEALTHCARE

 

Despite various challenges, limitations and setbacks, AI is virtually announced in healthcare sector. Below are some of the remarkable and revolutionary AI innovations in healthcare and pandemic era so far demonstrating a definite work in progress [9,10]:

Conclusion

Leveraging AI clinical decision support and risk assessment can be one of the most promising areas of development for bringing a revolution in health care services. Soon, AI will develop in a new era of clinical quality and exciting breakthroughs in patient care, by empowering a new generation of algorithms and tools. It will make clinicians more aware of pathological nuances and will make them more efficient while delivering care in a global pandemic like COVID-19. In no time, one will envision a world where a healthcare provider will enter an examination room, complete a diagnosis, explore treatment options, and then walk out of the room with discrete clinical data documented, orders placed, notes completed and charges captured.

 

Sources

 

 

1. Leslie Kane. Medscape National Physician Burnout & Suicide Report 2020: The Generational Divide. Medscape. 2020. Available at: https://www.medscape.com/slideshow/2020-lifestyle-burnout-6012460

2. Jones KE, Patel NG, Levy MA, Storeygard A, Balk D, Gittleman JL, et al. Global trends in emerging infectious diseases. Nature. 2008;451(7181):990–3.

3. Wiens J, Shenoy ES. Machine Learning for Healthcare: On the verge of a major shift in healthcare epidemiology. Clin. Infect Dis. Off Publ. Infect Dis. Soc. Am. 2018 Jan 1;66(1):149–53.

4. Jiang F, Jiang Y, Zhi H, Dong Y, Li H, Ma S, Wang Y, Dong Q, Shen H, Wang Y. Artificial intelligence in healthcare: past, present and future. Stroke Vasc Neurol. 2017 Jun 21;2(4):230-243. doi: 10.1136/svn-2017-000101.

5. Meskó, B., Görög, M. A short guide for medical professionals in the era of artificial intelligence. npj Digit. Med. 3, 126 (2020). https://doi.org/10.1038/s41746-020-00333-z.

6. Bansal A, Padappayil R.P, Garg C. et al. Utility of Artificial Intelligence Amidst the COVID 19 Pandemic: A Review. J Med Syst 44, 156 (2020). https://doi.org/10.1007/s10916-020-01617-3.

7. Vaishya R, Javaid M, Khan IH, Haleem A. Artificial Intelligence (AI) applications for COVID-19 pandemic. Diabetes Metab Syndr. 2020 Jul-Aug;14(4):337-339. doi: 10.1016/j.dsx.2020.04.012.

8. AI In Healthcare – Benefits, Challenges & Risks. Insight Brief. Available at: https://www.insightbrief.net/wp-content/uploads/AI-in-Healthcare-Benefits-Challenges-Risks-InsightBrief.pdf?tagged=true

9. Codrin Arsene. Artificial Intelligence in Healthcare: the future is amazing. Healthcrae weekly. 2020. Available at: https://healthcareweekly.com/artificial-intelligence-in-healthcare/

10. Sathian Dananjayan et al. Artificial Intelligence during a pandemic: The COVID‐19 example. Int J Health Manag. 2020; 35 (5): 1260-1262.